Statistical Physics Seminar

Michele Caraglio, Università di Padova, Italy

The many daily realizations of high frequency asset returns can be regarded as constituting an ensemble of histories amenable to statistical analysis. Taking into consideration the S&P 500 index, the scaling properties of the aggregated returns distribution can be employed to define a martingale stochastic model which consistently replicates the statistical properties of high frequency data in each trading day.
After presenting this model, I will show how it allows to construct density forecasts that can be used to implement a trend-following trading strategy capable of exploiting the small linear correlations present in the dataset.


In the second part of the talk I will deal with the case of a single time series of daily returns and I will give a sketch of how similar ideas regarding the scaling properties of returns permit to draw another model which aim at describing the dynamics of such a series and at reproducing many of its stylized facts.